2011-02-14 00:11 kristianpaul: http://downloads.qi-hardware.com/people/werner/tmp/teaser0.png http://downloads.qi-hardware.com/people/werner/tmp/teaser1.png http://downloads.qi-hardware.com/people/werner/tmp/teaser2.png http://downloads.qi-hardware.com/people/werner/tmp/teaser3.png 2011-02-14 00:12 wpwrak: teaser3 is done in your scope as well??? 2011-02-14 00:13 kristianpaul: with the data from the scope 2011-02-14 00:13 how how !! 2011-02-14 00:13 looks amazing clever 2011-02-14 00:14 lemme check in the scripts ... 2011-02-14 00:17 what is your scope brand?, i dint dig yet what remote programing or data acquisition can i do witht the tektronics 2011-02-14 00:17 i have a rigol ds1102cd 2011-02-14 00:24 kristianpaul: stores up to 1 MSa. or in this case (two channels) 512 kSa of each channel. unfortunately, your has much less memory. yours is a lot better for analog signals, though. 2011-02-14 00:26 hmm 2011-02-14 00:27 i', close to port sump to my avnet board so that will help just in case i need _just_ logic analizer thing 2011-02-14 00:30 kristianpaul: what was your model again ? 2011-02-14 00:31 wpwrak: TDS3012B, no digital Channels tought 2011-02-14 00:33 10 kSa memory ... should be enough for a single USB packet if you have sufficiently stable timing of the trigger 2011-02-14 00:34 the digital channels on the rigol don't work very well. i think i'd rather have more analog channels. 2011-02-14 00:36 of course, if you have something like MMC/SD to debug, then you're thankful for the digital channels. even if they have problems. (the problem with the rigol are that there are lots of glitches on the digital channels. you need to post-process them before they make much sense.) 2011-02-14 01:33 [commit] Lars-Peter Clausen: Restore sdio fix for the spectec wlan card http://qi-hw.com/p/openwrt-xburst/fdcac17 2011-02-14 01:37 larsc: thanks. (sorry for lose this patch) 2011-02-14 01:37 no problem 2011-02-14 01:39 hapylly testing his avnet sump logic analizer 2011-02-14 02:02 [commit] Werner Almesberger: atusb/fw3/an/: USB debugging scripts http://qi-hw.com/p/ben-wpan/f4d299d 2011-02-14 02:03 kristianpaul: here are my secrets weapons: http://projects.qi-hardware.com/index.php/p/ben-wpan/source/tree/master/atusb/fw3/an 2011-02-14 02:09 wpwrak: zprobe is from scope? 2011-02-14 02:12 zprobe: http://projects.qi-hardware.com/index.php/p/wernermisc/source/tree/master/zprobe 2011-02-14 02:12 ah misc 2011-02-14 02:12 i dont have that 2011-02-14 02:12 wait 2011-02-14 02:32 hates atmel's convention of expressing register bits in C by their position, not their mask. something like three days wasted chasing an addr | ADDEN vs. addr | 1 << ADDEN :-((( 2011-02-14 02:44 wpwrak: I was hunting a but without success the last 3 days. So if you had some succes then you should be happy 2011-02-14 02:44 I am not :( 2011-02-14 02:45 :) seems that my little stack works now. it even accepted the transmission of a 64+ bytes packet, which is what blow the freakusb stack apart 2011-02-14 02:46 (didn't check if it actually sent anything, though ;-) 2011-02-14 02:47 s/blow/blew/ 2011-02-14 02:49 [commit] Werner Almesberger: atusb/fw3/: adaptation of the f32xbase USB stack http://qi-hw.com/p/ben-wpan/da6677b 2011-02-14 02:51 zrafa: maybe you should try to debug usb. then you'd have your bug killed within 3 days, too ;-) 2011-02-14 02:54 yeah, transmission is good, too 2011-02-14 02:56 wpwrak: (works now) nice ;) 2011-02-14 02:56 one bug left, though. kernel complains that device descriptor read/64, error -71 (on the next try, it succeeds) 2011-02-14 02:59 wee ! 2011-02-14 03:00 wpwrak: http://downloads.qi-hardware.com/people/kristianpaul/sump.png 2011-02-14 03:00 guess what it is ;) 2011-02-14 03:00 argg the png was not fully exported 2011-02-14 03:00 kristianpaul: what program is that? 2011-02-14 03:01 kristianpaul: what hardware is that ? 2011-02-14 03:01 larsc: http://www.lxtreme.nl/ols/ 2011-02-14 03:01 wpwrak: sige evb > avnet (with sump bitstrem) > ols > png 2011-02-14 03:03 and what would avnet be ? 2011-02-14 03:03 oh sorry, avnet spartan 3e board 2011-02-14 03:03 wpwrak: http://ur1.ca/2jlp6 2011-02-14 03:04 aha ! :) 2011-02-14 03:04 that leads to the customer login ;-) 2011-02-14 03:04 ah? 2011-02-14 03:04 anyway, i can imagine what it is 2011-02-14 03:05 what's the sample rate ? how many samples can you store ? and is there any compression ? 2011-02-14 03:05 no idea 2011-02-14 03:05 I just got it work some minutes ago, let me check 2011-02-14 03:05 http://www.xilinx.com/products/devkits/aes_sp3a_eval400_avnet.htm 2011-02-14 03:06 256KSamples memory 2011-02-14 03:06 oh, that's a cheap board. nice. 2011-02-14 03:07 yes, it was my first touch with milkymist in hw 2011-02-14 03:08 ok, i got a 61.2 us sample 2011-02-14 03:08 speed is fixed to 115200 2011-02-14 03:09 compression i guess not 2011-02-14 03:09 so that's 256 kSamples that took 61.2 us ? 2011-02-14 03:11 that would be a sample rate of ~ 4 GSa/s 2011-02-14 03:11 s/us/ns 2011-02-14 03:12 hmm 2011-02-14 03:12 no, i'm cofused 2011-02-14 03:12 that on the other hand would be a meager 16 MSa/s 2011-02-14 03:12 sump.org/projects/analyzer/ <- more ingo 2011-02-14 03:13 inFo 2011-02-14 03:13 i still reading i think steve|m1 can quote something later, i new with this project 2011-02-14 03:14 (compression) that's something the fpga could implement. instead of storing fixed-clock samples, only store a sample when something has changed. in addition, store the number of non-changes since the last sample 2011-02-14 03:14 200 MHz sounds better :) 2011-02-14 03:16 nice it export data to text file 2011-02-14 03:17 100Msps captures up to 50MHz waveforms on 32 channels 2011-02-14 03:17 hmm i can get 8 channels with 24K sample depth 2011-02-14 03:18 maybe your board's characteristics differ from the DO-SPAR3-DK 2011-02-14 03:19 still spartan3, but yes 2011-02-14 03:19 you could port it to the SIE :P 2011-02-14 03:20 the DO-SPAR3-DK seems to have SRAM while the AES-SP3A-EVAL400-G doesn't 2011-02-14 03:20 you should go sleep :) 2011-02-14 03:21 so you'd be limited to 368640 memory bits, if digi-key's spec summary is to be trusted 2011-02-14 03:21 yup. xilinx say 360 kbits 2011-02-14 03:22 (block RAM) 2011-02-14 03:22 steve|m: may be, if jlime can run the osl, 2011-02-14 03:22 but no time now, i got my dummy logic analizer, i'm happy 2011-02-14 03:23 steve|m: you have a sie? 2011-02-14 03:23 it has tons of flash, though. so you could use that ;-))) 2011-02-14 03:24 yeah thats true 2011-02-14 03:26 kristianpaul: nope, but that would actually be a cool toy.. I like the idea of having the power of a fpga wired to a cpu 2011-02-14 03:26 is there a planned second run? 2011-02-14 03:27 to analyze USB, you'd need about 50 MSa/s. so that should be within the range of the board's capabilities 2011-02-14 03:27 i dont think so... but who knows, but i'm sure if you ask some of the carlos studet may sell one 2011-02-14 03:27 wpwrak: i cant usb now, as i lack buffer for 5v input.. 2011-02-14 03:28 isn't USB 3.3V signal level? 2011-02-14 03:28 the OLS has a buffer :P 2011-02-14 03:29 ah is 3.3V? 2011-02-14 03:29 heh 2011-02-14 03:29 xilinx is now actually bulding a fpga with an arm core the same chip 2011-02-14 03:29 with an intelligent trigger, you could just find the "start" bit and then immediately sample the packet. well, if you do intelligent clock recovery, you can just sample at the USB clock, 12 MSa/s 2011-02-14 03:29 USB D+/D- are 3.3 V 2011-02-14 03:30 good :-) 2011-02-14 03:31 larsc: spartan7? 2011-02-14 03:31 uhm, don't know exactly 2011-02-14 03:32 but they resently submitted support for it to lkml 2011-02-14 03:32 with 360 kbits of RAM, you could thus record about 22 kB of USB data or ~350 max-sized packets. that's already a reasonable analyzer 2011-02-14 03:35 I wonder if the whole on-chip ram is implemented.. 2011-02-14 03:35 well, i'm off, read you later 2011-02-14 05:15 [commit] Werner Almesberger: atusb: slightly adjusted some component values for easier sourcing http://qi-hw.com/p/ben-wpan/138a1be 2011-02-14 05:15 [commit] Werner Almesberger: atusb.brd: update and silk screen cleanup http://qi-hw.com/p/ben-wpan/fa99266 2011-02-14 08:20 hi all channel. 2011-02-14 08:22 wpwrak wolfspraul what you think on the last UBB mail with the price list, and preorder aproach time to announce on list? 2011-02-14 08:40 when i do make -j2 in openwrt build root, does it actually build a package in two threads or it builds two packages concurrently? 2011-02-14 08:41 it compiles multiple packages concurrently each with a single thread 2011-02-14 08:41 we cannot pass down -j by default as many packages are simply not safe to compile multithreaded 2011-02-14 08:41 trunk has black- & whitelisting for that and an additional feature to enable parallel build within packages 2011-02-14 08:41 sure, i just noticed an interesting thing 2011-02-14 08:42 according to make -j2 output, qt4 is already built (i.e. i see that next pacakges started to build) 2011-02-14 08:42 right now trying to benchmark OE with different bitbake threads/make threads numbers 2011-02-14 08:43 but in ps i can see that it is still building qt4 2011-02-14 08:43 kyak: grep PARALLEL_MAKE recipes/* in OE may show you recipes that have -j disabled :) 2011-02-14 08:43 so, you may just use our results :) 2011-02-14 08:44 kyak: maybe host vs target build of qt4 2011-02-14 08:44 Jay7: yes, there is NOPARALLEL or something like this in openwrt, too 2011-02-14 08:45 xMff: the output is currently at "keymouse" package 2011-02-14 08:45 i see this: "make[3] -C /home/bas/build/openwrt-packages/keymouse compile" 2011-02-14 08:45 I never messed with qt4 but afaik it is compiled for the host since qmake (?) is needed for other stuff 2011-02-14 08:45 and "make[3] -C feeds/packages/Xorg/lib/qt4 compile" is two lines above 2011-02-14 08:46 seems like finished, two other packages were built after that 2011-02-14 08:46 *shrug* :) 2011-02-14 08:46 :) 2011-02-14 08:46 i'm just wondering if -j2 can mess up things badly 2011-02-14 08:46 i didn't really try it before 2011-02-14 08:47 tuxbrain: you know me. preorder = fail. 2011-02-14 08:47 though i benchmarked, too - it takes around 30% faster 2011-02-14 08:47 if you totally have to do it, then try to limit the preorders to few large preorders, not many small ones. 2011-02-14 08:47 usually not, some time ago it used to fail in random places but nothing thata repeated make couldn't fix 2011-02-14 08:47 that way the communication overhead in case of unforeseen issues (read: always) stays managable 2011-02-14 08:47 so it at least does not result in catastrophic miscompilations 2011-02-14 08:48 ok :) 2011-02-14 08:48 most issues are missing inter-package dependencies that are never spotted with a single threaded make 2011-02-14 08:49 though some people invested quite some (cpu) time to iron those out 2011-02-14 08:52 would be good to have an auto build bot, like in some linux distros.. It would build a package (each package, or every new/updated package) in a separate environment (minimal build). It would reveal missing dependencies and help find other errors, too 2011-02-14 08:52 we have a buildbot farm but it is working single threaded atm 2011-02-14 08:53 I use it to keep track of broken packages 2011-02-14 08:53 http://buildbot.openwrt.org:8010/broken_packages/ 2011-02-14 08:54 cool! 2011-02-14 08:54 qt4 is broken in xburst :) 2011-02-14 08:55 there are false positives 2011-02-14 08:55 errors depend however greatly on build host (for example, i've been failing to build ruby for a long time because i have openssl-1.0.0) or other settings - like, usage of locale in uClibc or like different variants of iconv 2011-02-14 08:55 bah.., 3M buildlog for qt4... everything about this thing is big 2011-02-14 08:56 yup, qt4 is huge.. might be the longest package to compile 2011-02-14 08:57 hm, vim also builds fine.. what's the reason for these false positives? 2011-02-14 08:58 "make[#]: *** [...] Error ..." 2011-02-14 08:59 its basically an rgrep over the buildlogs 2011-02-14 08:59 the build itself runs with IGNORE_ERRORS=m so broken stuff is just skipped 2011-02-14 09:00 I have to go to wrok for once... bbl 2011-02-14 09:00 see you later 2011-02-14 09:17 larsc: about those fpga+arm cores. yes, that is not surprising and there is a whole wave of those types of chips going through the fpga landscape. 2011-02-14 09:17 I have 2 thoughts on this right now: 2011-02-14 09:18 1) some of those chips are priced very high, because everybody is testing how much value they create, and for whom, and who is in a stronger position, the FPGA side or the ARM/Power side. 2011-02-14 09:18 so if the chip is 90 USD or more, that's a big thing to consider 2011-02-14 09:18 also volumes are low, and right now it's just a test (there have been similar Power/fpga combos for a while already, and supported in Linux, but very expensive) 2011-02-14 09:19 2) I am very pragmatic on all this. If we compare a 'pure' FPGA solution running a free instruction set to a combo, we should pick whatever leads to an actually working powerful product faster. 2011-02-14 09:20 In any case I will only look at the free part of the solution, and try to achieve continuity there. So a solution like Milkymist SoC is more promising and flexible. But if a chip turns up with a great ARM core, and the core and overall system integration is easily targetable with free software tools, I would not hesitate to use it. Of course to advance the state of the free stuff still :-) 2011-02-14 09:21 hope this is not too vague. I keep an eye on the combos, but I feel very good with the Spartan-6 path of Milkymist and Milkymist One right now. 2011-02-14 09:22 [commit] Adam Wang: High speed bug fixed by Yanjun Luo: http://qi-hw.com/p/m1/f24e496 2011-02-14 09:52 [commit] kyak: supertux: don't install backup images http://qi-hw.com/p/openwrt-packages/d89a29e 2011-02-14 11:18 tuxbrain: replied by mail 2011-02-14 11:20 tuxbrain: regarding the preorders, they do indeed seem to add just unnecessary overhead. the money on the table would be EUR 450, which doesn't appear all that much. 2011-02-14 11:22 tuxbrain: preorders would make more sense if you were planning to make 1000+ units right away, and need firm commitments for that. this in turn would only matter if your pcb fab would charge you extra for 500 now and 500 later, as opposed to 500 now. that's often not the case. 2011-02-14 11:22 err, s/to 500 now/to 1000 now/ 2011-02-14 11:24 ok I will take in consideation that non-preorder aproach... (me crossing my fingers and put tap in the ears to not hear Victor complains about money....) 2011-02-14 11:26 tuxbrain: it would make sense to do 500 units no matter what sales projection you have. (well, unless it's 0 ;-) even if someone wanted to buy a few million pieces, you'd still want to have a "small" first run to see if things are alright, before committing to more. 2011-02-14 11:35 There are sites where you can order cheap pcb's for self designed open source hardware.Â Â Are there places where you can have smd chips soldered on the board, or even build the whole product? 2011-02-14 11:36 Without a huge investment? 2011-02-14 11:40 kuribas: if you had omitted the second message, i could have suggested some ;-)) 2011-02-14 11:41 kuribas: well, it all depends on what "huge" means to you. for the pcb, is a fixed setup cost of about USD 500 a lot ? not sure what the setup cost for SMT would be. probably at least USD 1000. 2011-02-14 11:43 viric: you sound as if sourcing wasn't already a story co-autored by kafka, de sade, and dante ;-) 2011-02-14 11:46 [commit] kyak: terminus-font: initial port http://qi-hw.com/p/openwrt-packages/a14421b 2011-02-14 11:46 wpwrak: I guess that's very reasonable, but for me personally it would be a lot.Â Â There was a page from seeedstudio where you could make 10 pcb's for only 50Â¬ without startup cost, if you make the design open source. 2011-02-14 11:47 Maybe it's better if I started working on my smd skills :-) 2011-02-14 11:47 smd soldering skills I mean. 2011-02-14 11:48 When I look at the projects on sparkfun and similar sites, they are mostly build on arduino and robotics.Â Â There is very little pro audio. 2011-02-14 11:48 wpwrak: uh? What is it about? :) 2011-02-14 11:49 wpwrak: I don't understand "sourcing" 2011-02-14 11:49 hm I got disconnected 2011-02-14 11:49 wpwrak: what is that 'sourcing' about? I don't understand :) 2011-02-14 11:50 viric: sourcing is the process of selecting and acquiring the components you need for production 2011-02-14 11:50 ahh 2011-02-14 11:50 viric: well, the selection could be considered apart. but in practice, the two are intimately connectd 2011-02-14 11:51 sure 2011-02-14 11:51 kuribas: (10 PCBs for 50 EUR) sounds like a great deal ! there are also places like batchpcb, which have low cost. (but limited technology and long turn-around times) 2011-02-14 11:51 Well, I would not like to know that the Ben Nanonote pieces are produced in factories with a 'high score' on violating human rights :) 2011-02-14 11:52 or things like forbidding unions and things like that 2011-02-14 11:52 viric: then you probably shouldn't try to find out ;-) i wouldn't expect any particular atrocities, but you never know 2011-02-14 11:53 viric: hmm, in some places of the world, disallowing unions sounds like a most excellent idea 2011-02-14 11:53 viric: (that is, if you can, because there they're pretty much equivalent to the mafia) 2011-02-14 11:53 wpwrak: It does.Â Â I wonder if the same exists for smd soldering.Â Â The problem is that the adc and controller I need are smd.Â Â If I would sell a kit, I'd probably make a mess trying to solder those components ... 2011-02-14 11:55 kuribas: smd is often easier to solder than through-hole. what packages are you talking about for the larger chips ? ssop/tsop/sioc/etc. are trivial. qfn is quite doable if you don't need the center pad to be 100% reliable. bga and friends would be difficult for manual soldering. 2011-02-14 11:56 wpwrak: that sounds like a usual neoliberal excuse :) 2011-02-14 11:56 let's see 2011-02-14 11:57 viric: i would think less of the intelligence of the neoliberals if they wouldn't use such glaring examples of abuse of power for their agenda, too :) 2011-02-14 11:57 wpwrak: they seem to work 2011-02-14 11:58 wpwrak: On doubt, on contradictory reports from the workers union and the company owners, I go on trusting the workers .) 2011-02-14 11:58 :) 2011-02-14 11:59 viric: oh, sure they work. they even make their employees earn a lot. e.g., a truck driver in argentina earns more than a pretty much anyone with a phd. the whole transportation system is also horribly inefficient, dangerous, and environmentally hostile. 2011-02-14 12:00 wpwrak: pcm1803 is 20 pin ssop, at19sam7 is qfn or lqfn. 2011-02-14 12:00 viric: let alone unreliable. dangerous = argentina has one of the highest if not the highest rates of road deaths. and in a lot, trucks are involved. 2011-02-14 12:01 wpwrak: here, phd students rarely try to do any collective effort to negotiate their working conditions, too 2011-02-14 12:01 viric: of course, whenever the idea of replacing long-distance transportation with, say, trains comes up, the trucker union start pressuring. 2011-02-14 12:02 wpwrak: it has become "trendy" not to do any collective effort to negotiate anything. And it's also trendy to untrust the unions, and their intentions of getting good working conditions for their members. 2011-02-14 12:02 wpwrak: ah, sure, this also happens here 2011-02-14 12:02 kuribas: the arm may be a bit difficult. the ssop is no problem. 2011-02-14 12:03 wpwrak: Do you use any special equipment for soldering smd?Â Â A solder iron with temperature control? 2011-02-14 12:05 wpwrak: what did the union of executioners do in Argentina, when they saw their jobs on peril? :) 2011-02-14 12:06 viric: i would rather look for the flaws in the political system. this kind of unions is quite anti-democratic yet wields power comparable to that of the government. doesn't this make you worry a little ? 2011-02-14 12:06 wpwrak: the political system is biased towards the creditors of the political parties 2011-02-14 12:07 viric: executioners ? huh, attack some islands ? :) 2011-02-14 12:07 viric: there you have your problem 2011-02-14 12:07 wpwrak: and the first priority of the political class is to perpetuate the political class :) 2011-02-14 12:07 (lunch!) 2011-02-14 12:07 kuribas: yup. a decent soldering iron, flux, and thin solder 2011-02-14 12:08 wpwrak: I know the problem. But those situations have to come into light. And when there is the option to buy one product or another, we have our vote when choosing 2011-02-14 12:08 viric: for affecting working conditions, you're at the wrong place here. nothing sharism could do would have even the slightest effect. 2011-02-14 12:09 viric: so it would purely be to make you feel righteous, but you'd lie to yourself if you thought you had improved anything 2011-02-14 12:10 viric: if you want to make an impact, find the big buyers, analyze their sources, and then build public pressure if you find anything there 2011-02-14 12:10 viric: if you can make apple, htc, samsung, nokia, sony, etc., impose requirements on their suppliers, then you have a chance of this to have an effect 2011-02-14 12:12 viric: but small companies don't matter. lots of vendors wouldn't even talk to small operations like sharism because their volume is insignificant. 2011-02-14 12:15 viric: small companies basically take the crumbs of what the big ones had. so you really have to aim for the big ones. or become big yourself :) 2011-02-14 12:28 tuxbrain: be careful about 'no preorder'. you may very well end up with a lot of boards. 2011-02-14 12:28 my point was more subtle 2011-02-14 12:28 on one hand preorders are good, they tie people into the project early 2011-02-14 12:29 make them prepay 50%! not 100%, not 0% 2011-02-14 12:29 on the other hand preorders cause communication and admin overhead that is often overlooked. that is the more so true the more innovative/new/risky a project is. 2011-02-14 12:29 because - if something unexpected happens (no need to give an example, you know the list is infinite), then you need to communicate that back to all preorderers 2011-02-14 12:30 and you need to give them reaction options a) b) c) 2011-02-14 12:30 you need to collect feedback 2011-02-14 12:30 etc. etc. 2011-02-14 12:30 it's terrible 2011-02-14 12:30 wolfspraul: naw, make them prepay 100%. we're talking about quite trivial amounts. 2011-02-14 12:30 so you should go for preorders, but you need to deal with at most 5 people or so 2011-02-14 12:30 wpwrak: I think large preorders are ideal, because of communication overhead as described. 2011-02-14 12:31 that's a good mix between putting the risk on multiple shoulders, while keeping everybody fully informed 2011-02-14 12:31 wolfspraul: don't you think the first run should be small anyway ? 500, maximum 1000 units. see if anything goes bad. 2011-02-14 12:31 so if tuxbrain would have 4 people he knows will take 100 each, that's ideal 2011-02-14 12:31 100 for himself 2011-02-14 12:31 yes sure 2011-02-14 12:31 wolfspraul: yup. that sounds reasonable. 2011-02-14 12:31 I just share my preorder experience. 2011-02-14 12:32 it can be good, it can backfire. 2011-02-14 12:32 so my advice is: 1) deal only with a small number of large preorderers, let's say 5-10. 2) make them prepay 50%, that will keep everybody honest and committed. 2011-02-14 12:32 wolfspraul: if preorders are really necessary, i wouldn't split them. not on such small values and with business partners who largely know each other. 2011-02-14 12:32 if you can satisfy those two requirements, that's much better than taking the entire risk yourself 2011-02-14 12:38 i mean, it's not like sending a few kbucks to Happy No-Name Electronics and Rice Trading Ltd., Ã†â€žÃ¢â€”Q 2011-02-14 12:45 or Mugu Intl. Enterprises, South Cammer, Nigeria ;-) 2011-02-14 12:47 (jtag) heh, so close ;-) 2011-02-14 13:09 kristianpaul: Adam said you fixed your jtag-serial high-speed problem already? 2011-02-14 13:12 cool, Adam uploaded really nice rework instructions. with those pictures I think even I could try :-) 2011-02-14 13:12 http://en.qi-hardware.com/wiki/JTAG_Serial_Cable_run_1_for_Milkymist_One#Bugs,_Steps_of_Rework_and_Improvements 2011-02-14 13:34 wpwrak: well, among certain public, having some good labels about good procedures and decisions can result in benefits 2011-02-14 13:34 even for the company 2011-02-14 13:36 wpwrak: I'm not trying to accuse qi of anything :) I just wondered, if part of the 'all free' politics went further in chosing vendors 2011-02-14 13:37 viric: I'm all with you. the deeper the knowledge the better, right? 2011-02-14 13:37 yes, I would like to know about social standards at component and part vendors. 2011-02-14 13:37 viric: i think general egalitarian principles certainly resonate well with the general objectives 2011-02-14 13:38 but we have to be realistic. collecting this kind of data, especially if it should be accurate, is terribly hard and time consuming work. 2011-02-14 13:38 the only way I can see this happen is if it's some kind of wiki-style collective process. 2011-02-14 13:38 wikivendors.org 2011-02-14 13:38 a global initiative to collect all sorts of data points about every company on the planet :-) 2011-02-14 13:39 nicely meta-tagged, searchable, updated by 100,000 volunteers, by insiders of the companies (ideally in an official function and not as whistleblowers), and so on 2011-02-14 13:39 please don't expect that the tiny Qi outfit can make any serious headway there. 2011-02-14 13:39 I try my best, with the few working hours I have each day. 2011-02-14 13:40 I like sourcemap.org, for example. at least it's a project that seems to point in the right direction. 2011-02-14 13:40 wolfspraul: correctly assessing social standards may be pretty hard. at least if you want to go beyond just western imperialism :) (like demand that chinese factory worker gets the equivalent of an average US meal ;-) 2011-02-14 13:40 actually I believe it is totally doable 2011-02-14 13:40 the world is not as black and white as we sometimes make it 2011-02-14 13:41 but collecting the data on a timely basis, and keeping it accurate - it can only happen as a big collective project 2011-02-14 13:41 no one vendor can do it 2011-02-14 13:41 I don't like the buy a product if I would know they put some 12-year old to work. 2011-02-14 13:41 definitely not 2011-02-14 13:42 wolfspraul: doable yes. but much harder than just collecting some statistics. you probably need to spend months on analyzing each location. establish local sources. validate them. and so on. 2011-02-14 13:42 we need to pull out this kind of information. but then it's not the core mission of Qi, at best we can join some other projects working in that direction. 2011-02-14 13:42 oh sure. that's what I'm saying. 2011-02-14 13:42 wolfspraul: I'm a subscriber to some people that investigate how production works. Every two month they analise some kind of production. Some moths ago they investigated the mobile phone producer. 2011-02-14 13:42 it needs to be a huge collective project. 2011-02-14 13:42 s 2011-02-14 13:43 wolfspraul: of course they analised only major brands 2011-02-14 13:43 that sounds good. I am 'investigating' dozens of factories for years :-) 2011-02-14 13:43 there are many worthwhile things one could improve there. 2011-02-14 13:43 soldering exhaust fumes for example 2011-02-14 13:43 really bad sometimes! 2011-02-14 13:44 coughs 2011-02-14 13:44 dumping chemical waste into the next river 2011-02-14 13:44 excessive over-hours (80hr/week) 2011-02-14 13:44 I have never seen child-labor, in China that is really rare and definitely not happening in the core technology industry. 2011-02-14 13:44 bah, lavish spare time :) 2011-02-14 13:44 exactly 2011-02-14 13:44 maybe it's more common in south-east asia or non-tech industries. 2011-02-14 13:45 that sourcemap looks good. I did not know it. 2011-02-14 13:45 yeah 2011-02-14 13:45 I like it too. 2011-02-14 13:45 but so far my participation hasn't evolved beyond 'I like it' :-) 2011-02-14 13:45 back to reality... 2011-02-14 13:45 :) 2011-02-14 13:46 You mean "keeping on ignoring reality" :) 2011-02-14 13:46 (temporarily) 2011-02-14 13:46 no back to the reality of how hard it is to collect this kind of data 2011-02-14 13:47 sure 2011-02-14 13:47 "How easy it is to get the information" is one of the score measures, for companies. 2011-02-14 13:47 actually many people would like to participate 2011-02-14 13:47 I wouldn't worry about that. 2011-02-14 13:47 also on the vendors side 2011-02-14 13:47 in fact you could easily differentiate and label between participating and non-participating vendors 2011-02-14 13:47 exactly 2011-02-14 13:48 where 'participation' is defined by certain standards, like response times, dedicated contact, etc. 2011-02-14 13:48 but it's still a huge amoutn of work 2011-02-14 13:48 and then you need enough critical mass 2011-02-14 13:48 (as always for new things) 2011-02-14 13:48 I cannot do much, except 'liking' sourcemap.org 2011-02-14 13:48 I was not requiring anything like that to the few qi workers. I simply had hope that in the ben nanonote world it would be easier to find people with such worries; easier than in the Macintosh world :) 2011-02-14 13:49 oh totally 2011-02-14 13:49 I am with you. 2011-02-14 13:49 and we try to document vendors 2011-02-14 13:49 but even then each vendor is still a big black box. 2011-02-14 13:49 Clear. 2011-02-14 13:50 we had a good documentary here... "programmed obsolescence" 2011-02-14 13:51 there was a good part about http://www.greenpeace.org/international/en/news/features/poisoning-the-poor-electroni/ 2011-02-14 13:54 (and of course, a lot about designing the devices to last short) 2011-02-14 13:54 (is 'last short' the opposite of 'last long'?) 2011-02-14 13:54 unfortunately there are strong economic incentives for short product cycles 2011-02-14 13:54 short product cycles drive up volume, and the higher volume drives down unit prices 2011-02-14 13:55 that makes a strong recycling program all the more important, and one needs to say that Apple (the little I know about recycling) is leading there as well, basically they are trying to get every owner of say an iPad to always come back to a store if it's broken, and return it there) 2011-02-14 13:56 I read about a program Apple has where if you bring your old broken iPad to the store, they give you a new one for 99 USD? I think the types of defects that would quality were listed. 2011-02-14 13:57 anyway this leads in the right direction, products should return the same path they came from, because that's where the most knowledge about the products is located 2011-02-14 13:58 or alternatively products should be released with full and deep documenation, then it doesn't matter in which hands they end up eventually, those people can always easily (that's the point) look it up 2011-02-14 13:58 we are far from those noble goals right now, including Ben NanoNote or Milkymist One... 2011-02-14 14:00 I can relate to that ghana video, there's a lot of such places and the health situation of neighboring villages is the least concern of the owners of such places. so a brand manufacturer has the reponsibility to keep its products away from such irresponsible people. 2011-02-14 14:00 easier said than done though 2011-02-14 14:01 in China you regularly read news of a thousand kids suddenly loosing their teeth in some village, because some dubious factory in or around the village did some dubious thing that poisoned drinking water. not good. 2011-02-14 14:02 wolfspraul: what happens with the people responsible for this ? nothing ? 2011-02-14 14:02 to be fair these companies are real bottom feeders and if you buy products from any brand company, they will by far stay away from such outfits. 2011-02-14 14:03 no, if the news comes out they will be punished severely. 2011-02-14 14:03 just keep in mind they are totally uneducated 2011-02-14 14:03 so the punishment hits them like a lightning 2011-02-14 14:03 good. so there's at least an incentive 2011-02-14 14:03 they have no clue what they did, or why the kids lost their teeth 2011-02-14 14:03 not really 2011-02-14 14:03 wolfspraul: I know quite well about the 'economic incentives'. 2011-02-14 14:03 they are far too uneducated to anticipate the punishment 2011-02-14 14:03 (uneducated) hmm, i see 2011-02-14 14:04 wolfspraul: optimising for economics (meaning positive balance in the company accounting) does not match optimising many other things that end up becoming "irrelevant" or "annoying" to the optimisation, like even respecting human rights :) 2011-02-14 14:05 well, I think it's also a good thing. 2011-02-14 14:05 a low unit price is also important to make the product affordable. 2011-02-14 14:05 in hardware high volume helps bring down unit costs. 2011-02-14 14:05 so if you design for short product cycles, you just fulfill a requirement of your own industry. 2011-02-14 14:05 the question is what happens with the returned products. 2011-02-14 14:06 if they are properly handled, what should be bad about it? 2011-02-14 14:06 maybe there are theoretical limits to how perfect recycling can be? don't know 2011-02-14 14:06 assuming it can be perfect, I have no problem to 'throw away' my phone every month, if that means a phone just costs 5 USD :-) (throw away = fully recycled) 2011-02-14 14:07 someone would need to find out what's possible 2011-02-14 14:07 I always wnated to manufacture something out of used ICs, just never had the guts/time/money to try... 2011-02-14 14:07 probably it would end in disaster :-) 2011-02-14 14:07 viric: i think a healthy egoism is a good thing. that's a motivation i find perfectly credible and easy to understand :) 2011-02-14 14:08 viric: the difficulty lies in making sure people understand what's best for them 2011-02-14 14:09 totally agree 2011-02-14 14:17 wpwrak: there should be some people doing the 'reminding'. I think all the current situation goes on only because of ignorance. 2011-02-14 14:18 and people expect to get a computing device for $100 only because they have seen devices at that price, ignoring anything behind. 2011-02-14 14:18 viric: (ignorance) i don't think that's all. sometimes, you also need some concentration of consequences. like laws that are enforced. 2011-02-14 14:18 I don't think the problem is about telling the people what's best for them. It's only about keeping people aware. 2011-02-14 14:18 wpwrak: it's part of the things to be aware of 2011-02-14 14:20 wolfspraul: 'recycling' does not play a good role in the optimisation equations for companies, as long as waste can be thrown to "bought dictatorships offshore". 2011-02-14 14:20 I don't think it's that bad, really. 2011-02-14 14:20 good companies work hard to improve this 2011-02-14 14:20 otherwise our world would be a crazy place already 2011-02-14 14:20 wolfspraul: it's crazy in some places, not crazy on some others. 2011-02-14 14:20 the cases we are talking about are rare cases, and 90% or more of good companies work against this. 2011-02-14 14:20 viric: plus, you need to have enough to lose to prevent you from taking crazy risks. that's why i see the most important factors for a decent society in a "middle class" plus a fair and functional legal system. remove any of these, and things fall apart. 2011-02-14 14:21 the internet can definitely make everybody jump forward on this subject 2011-02-14 14:21 wikivendors.org 2011-02-14 14:21 think about how primitive things like ISO 9001 or 14001 are in comparison. 2011-02-14 14:21 it's a joke 2011-02-14 14:21 just some formal procedures 2011-02-14 14:22 but those things are installed and working, so fine 2011-02-14 14:22 the internet and ideas like wiki/collaborative editing will allow us to bring this to a whole new level 2011-02-14 14:22 so let's just start! :-) 2011-02-14 14:22 well, there are projects where to collaborate. We are not the first on this 2011-02-14 14:22 we (the consumers) are far more than a few producers, especially the even fewer bad/greedy/irresponsible producers 2011-02-14 14:23 it will be very easy to single them out, once an effective way of collaboration has been established 2011-02-14 14:23 yeah perfect, then just do it. and share the links of good projects here. 2011-02-14 14:23 As with many other things, it's the internet against the broadcast media 2011-02-14 14:25 As far as I find anything, I'll let know here as always :) 2011-02-14 14:25 of course I thought it was more productive writing here than writing a letter to Apple or Nokia :) 2011-02-14 14:26 don't underestimate those guys. give them the benefit of the doubt at least. 2011-02-14 14:26 they may have very responsive people working on this. 2011-02-14 14:27 from Apple in particular I can tell you they are going into their supply chain several levels deep. 2011-02-14 14:27 hm I've heard some different reports 2011-02-14 14:27 Apple sends Apple employees to remote places (suppliers of suppliers), where no other foreigner ever has been seen, except people from Apple and crazy Qi people 2011-02-14 14:28 but Apple for sure is not the worst around 2011-02-14 14:28